Thesis | Disturbance, vegetation co-occurrence, and human intervention as drivers of plant species distributions in the sagebrush steppe

Changes in fire regimes, invasive species dynamics, human land use, and drought conditions have shifted important plant species in the Northern Great Basin (NGB)— including big sagebrush (Artemisia tridentata ssp.), conifers (e.g., Juniperus spp.) and invasive annual grasses (e.g., Bromus tectorum). Characterizing how these overlapping disturbances influence species distributions is critical for land management decision- making. Previous research has explored the individual effects of drought, wildfire, restoration, and invasive species on sagebrush steppe communities, but the specific effects of these disturbances in context with one another remain poorly understood at a landscape scale. To address this gap, I constructed multilevel conditional autoregressive (CAR) species distribution models (SDMs) to map the distributions of big sagebrush, juniper, and cheatgrass on lands managed for grazing in the NGB, both with and without a history of fire. These models illuminate the concurrent influences of species co-occurrences, drought, wildfire characteristics (e.g., fire size, time since fire, and number of fires), and restoration treatments. For all SDMs, results indicate that species co-occurrence exhibits the strongest effect—between 1.23 and 19.2 times greater than the next strongest predictor—on all species’ probability of occurrence, suggesting that vegetation co-occurrence meaningfully influences landscape-scale species distributions. In portions of the NGB both with and without historical fire, number of fires and maximum vapor pressure deficit (VPD) also exert substantial influence on the likelihood of species presence, and results indicate that restoration treatments have broadly met desired outcomes for both sagebrush and juniper. Narrowing down to only areas that have previously burned, however, models do not support the efficacy of post-fire restoration. All versions of the SDMs, which rely on Bureau of Land Management-administered grazing allotments as a spatial varying intercept, also explicitly point to the differential influence of long-term management regimes on species distributions. These model predictions capture post-disturbance vegetation outcomes under changing fire, climate, and invasive species regimes and in the context of human decision-making, in turn defining a plausible ecological space as these disturbance and management processes play out into the future.